INDEX

Explanations

the word "reason"

oai_token-act-pair · gpt-3.5-turbo

phrases indicating causes or explanations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xon

-0.72

Appearance

-0.68

chio

-0.68

ccording

-0.67

ach

-0.66

wana

-0.65

ategor

-0.65

uron

-0.65

avin

-0.64

rons

-0.64

POSITIVE LOGITS

why

1.65

why

1.22

abl

1.19

WHY

1.17

 behind

1.04

Why

0.95

Why

0.83

 cited

0.83

for

0.78

 cele

0.74

Activations Density 0.036%

the word "reason"

phrases indicating causes or explanations

No Comments

No Known Activations